diff options
Diffstat (limited to '')
| -rw-r--r-- | new-luxc/source/luxc/parser.lux | 65 |
1 files changed, 19 insertions, 46 deletions
diff --git a/new-luxc/source/luxc/parser.lux b/new-luxc/source/luxc/parser.lux index 1e280e62b..7d9c77f2b 100644 --- a/new-luxc/source/luxc/parser.lux +++ b/new-luxc/source/luxc/parser.lux @@ -30,7 +30,6 @@ (lux (control monad ["p" parser "p/" Monad<Parser>]) (data [bool] - [char] [text] ["R" result] [number] @@ -160,29 +159,28 @@ ## and 4 characters long (e.g. \u12aB). ## Escaped characters may show up in Char and Text literals. (def: escaped-char^ - (l;Lexer [Text Char]) + (l;Lexer [Nat Text]) (p;after (l;this "\\") (do p;Monad<Parser> [code l;any] (case code ## Handle special cases. - "t" (wrap ["\\t" #"\t"]) - "v" (wrap ["\\v" #"\v"]) - "b" (wrap ["\\b" #"\b"]) - "n" (wrap ["\\n" #"\n"]) - "r" (wrap ["\\r" #"\r"]) - "f" (wrap ["\\f" #"\f"]) - "\"" (wrap ["\\\"" #"\""]) - "\\" (wrap ["\\\\" #"\\"]) + "t" (wrap [+2 "\t"]) + "v" (wrap [+2 "\v"]) + "b" (wrap [+2 "\b"]) + "n" (wrap [+2 "\n"]) + "r" (wrap [+2 "\r"]) + "f" (wrap [+2 "\f"]) + "\"" (wrap [+2 "\""]) + "\\" (wrap [+2 "\\"]) ## Handle unicode escapes. "u" (do p;Monad<Parser> - [code (l;between +1 +4 l;hex-digit)] - (wrap (case (:: number;Hex@Codec<Text,Nat> decode - (format "+" code)) + [code (l;between +1 +4 l;hexadecimal)] + (wrap (case (|> code (format "+") (:: number;Hex@Codec<Text,Nat> decode)) (#;Right value) - [(format "\\u" code) (char;char value)] + [(n.+ +2 (text;size code)) (text;from-code value)] _ (undefined)))) @@ -190,31 +188,17 @@ _ (p;fail (format "Invalid escaping syntax: " (%t code))))))) -## A character can be either a normal glyph, or a escaped character. -## The reason why this parser returns both the Char and it's textual -## representation in the source-code, is for the sake of updating the -## cursor after parsing the char. -## A character only represents one glyph, but it's source-code -## representation may be multi-glyph (e.g. \u1234, \n), in which case, -## the text that was parsed needs to be counted to update the cursor. -(def: raw-char^ - (l;Lexer [Text Char]) - (p;either (do p;Monad<Parser> - [char (l;none-of "\\\"\n")] - (wrap [char (|> char (text;nth +0) assume)])) - escaped-char^)) - ## These are very simple parsers that just cut chunks of text in ## specific shapes and then use decoders already present in the ## standard library to actually produce the values from the literals. (def: rich-digit (l;Lexer Text) - (p;either l;digit + (p;either l;decimal (p;after (l;this "_") (p/wrap "")))) (def: rich-digits^ (l;Lexer Text) - (l;seq l;digit + (l;seq l;decimal (l;some rich-digit))) (def: (marker^ token) @@ -262,16 +246,6 @@ number;Codec<Text,Deg>] ) -## This parser doesn't delegate the work of producing the value to a -## codec, since the raw-char^ parser already takes care of that magic. -(def: #export (parse-char where) - (-> Cursor (l;Lexer [Cursor Code])) - (do p;Monad<Parser> - [[chunk value] (l;enclosed ["#\"" "\""] - raw-char^)] - (wrap [(update@ #;column (|>. ($_ n.+ +3 (text;size chunk))) where) - [where (#;Char value)]]))) - ## This parser looks so complex because text in Lux can be multi-line ## and there are rules regarding how this is handled. (def: #export (parse-text where) @@ -334,10 +308,10 @@ ## Must handle escaped ## chars separately. (do @ - [[chunk char] escaped-char^] - (recur (format text-read (char;as-text char)) + [[chars-consumed char] escaped-char^] + (recur (format text-read char) (|> where - (update@ #;column (n.+ (text;size chunk)))) + (update@ #;column (n.+ chars-consumed))) false)) ## The text ends when it ## reaches the right-delimiter. @@ -538,15 +512,14 @@ (parse-deg where) (parse-symbol where) (parse-tag where) - (parse-char where) (parse-text where) ))) (def: #export (parse [where code]) (-> [Cursor Text] (R;Result [[Cursor Text] Code])) - (case (p;run code (parse-ast where)) + (case (p;run [+0 code] (parse-ast where)) (#R;Error error) (#R;Error error) - (#R;Success [remaining [where' output]]) + (#R;Success [[_ remaining] [where' output]]) (#R;Success [[where' remaining] output]))) |